Experimental Progress Zebrafish visual behaviors: We have made significant progress toward incorporating the larval zebrafish as a model system in the lab. Our short-term goal is to use a custom light sheet microscope to image nearly the entire brain of a larval zebrafish while the animal performs a variety of behaviors, but focusing on prey capture. We will eventually acquire 3D electron microscopy datasets from fish that we have previously imaged. Dr. Chris Harris (postdoc) established a mating protocol and incubator to raise fish from fertilization to 7 dpf in the lab. He then developed a closed loop visual stimulator to record the tail position using high speed videography while a fish views a virtual prey-like stimulus. The fish is partially restrained to stabilize the brain, but the eyes and tail are free to move. The tail position is used to update the location of a prey-like stimulus on the visual stimulator. Chris has successfully recorded orienting turns toward prey-like stimuli and is currently collecting behavioral data to determine how frequently fish respond to a stimulus. The light sheet microscope is nearing completion (see below) and we expect to begin imaging from transgenic behaving larval zebrafish in Oct. 2013. The pan-neuronal calcium indicator fish we will use are being generated in the lab of Dr. Harry Burgess (NICHD). Neurogenesis in the olfactory bulb: We are interested in the circuitry of the initial stages of olfactory processing in the olfactory bulb (OB). In addition, we are studying neurogenesis in this system, specifically the integration of perinatal and adult born granule cells in the circuitry of the OB. Our goal is to label perinatal and adult born granule cells with two different AAV viruses encoding two different color fluorescent proteins. Thus, color encodes the date of birth of the labeled granule cells. Then, we will optogenetically stimulate mitral and tufted cells in the OB while imaging calcium transients in the two populations of granule cells to determine whether there is any age-dependent difference in the functional connectivity. Finally, we will reconstruct the connectivity within an OB column by 3D electron microscopy. Dr. Marta Pallotto (postdoc) has successfully performed double survival surgeries, injecting wild-type mice in the rostral migratory stream at P2 with AAV-CFP and the same mice at P40 with AAV-RFP. She has imaged the OBs of these animals and confirmed two populations of granule cells each expressing one of the two colors, indicating that she is able to encode birth date with color. In parallel she has bred a transgenic mouse line expressing channelrhodopsin selectively in mitral and tufted cells. Marta is currently learning how to label the AAV expressing granule cells with a synthetic calcium indicator and how to record electrophysiologically from granule cells in the cell-attached mode (to calibrate calcium transients). She will soon begin optogenetically stimulating mitral and tufted cells while performing simultaneous calcium imaging in granule cells to test for functional connectivity. In parallel, she will begin collecting 3D EM datasets from the OB. Technical Progress Development of custom light sheet microscope: I designed and constructed a novel light sheet microscope to enable whole-brain imaging of larval zebrafish. The design is unique in that the imaging is performed upright instead of horizontally (as in commercial systems). This allows us to image from semi-restrained swimming fish. The microscope also uses two-photon instead of one-photon excitation to avoid excitation of retinal photoreceptors by the illumination light. The microscope incorporates the closed-loop visual stimulator that Chris Harris developed. Matt Schiel (summer intern) has written open-source software to control the microscope from Matlab. This microscope is currently being tested and will be in use within the next couple weeks. OLED-based optogenetic stimulator: Matt Schiel and I implemented a optogenetic stimulator that allows arbitrary temporal and spatial excitation of channelrhodopsin in acute brain slices. We integrated this stimulator into our existing two-photon laser scanning microscope. Marta Pallotto is currently using the stimulator to excite mitral and tufted cells in OB slices. Development of custom serial-block face microtome: I designed and constructed several prototypes for a serial block-face SEM microtome. The most recent version is capable of automatically sectioning specimens in our SEM vacuum chamber at 40nm section thickness. I am actively refining the design of this microtome to further reduce the section thickness. We have used the microtome to collect a number of test image stacks from both zebrafish and OB specimens. Novel EM staining: Christian Hicks (visiting medical student) joined us in July 2013 and has begun working on novel methods to improve the staining of tissue for SBEM. He is exploring several promising avenues with the goal of increasing the contrast between plasma membranes and intracellular organelles and cytosol. Christian is also working on fluorescently labeling long range axons to allow us to perform correlative light and EM microscopy. Quantum dot labeling: Boma Fubara (contractor, lab technician) has been attempting to develop a labeling protocol to increase the depth penetration of quantum dots. We wish to use quantum dots for correlative light and EM microscopy, but their usefulness is currently limited by a penetration depth of only a few microns into tissue sections. Boma is also working on immuno-EM methods to label the transmitter content of synaptic vesicles. 3D EM data analysis: Dr. Paul Watkins (contractor, software engineer) joined us in July 2013 to develop an image analysis pipeline for our large 3D EM datasets. He has already successfully implemented a machine learning algorithm based on convolutional neural networks to generate fully automated segmentations. His results are currently matching those presented at recent analysis challenges. He continues to optimize this approach to further reduce the test error from these networks and improve the segmentations. Web based Proofreading Tool: Christina Burghard (postbac IRTA) and Amauche Emenari (postbac IRTA) have worked in collaboration with the NINDS Bioinformatics Group to develop a web-based analysis tool. The tool allows us to collect training data and human annotations (corrections) of the automated neural network output from remote users. The beta version of this tool is nearly complete and we will begin testing it internally at NIH. Our goal is to deploy the tool externally by the end of 2013.